Introducing NVIDIA DGX Station for Windows, the world's most powerful deskside AI supercomputer with Windows powered by NVIDIA GB300. ✅ Run frontier AI models with up to 1 trillion parameters locally ✅ Build and run secure AI agents on Windows with NVIDIA OpenShell ✅ Built by @ASUS, @Dell, @GIGABYTE, @HP, @msigaming, and @Supermicro #NVIDIAGTC https://t.co/82tqvvNzJU
NVIDIA brings trillion parameter AI models to Windows enterprise desktops
- Model Capacity
- Up to 1 trillion parameters
- Compute Performance
- 20 petaflops FP4
- System Memory
- 748GB coherent memory
- Networking
- 800Gb/s ConnectX-8 SuperNIC
- Processor
- 72-core NVIDIA Grace CPU
Enterprise AI is shifting toward autonomous agents that need direct access to Windows applications. While heavy inference previously required Linux data centers, this hardware brings that compute to the desktop, joining the NVIDIA RTX Spark launch to scale local agentic performance. This move tracks alongside the Windows-centric focus of OpenAI's Windows update.
Users can build secure agents via NVIDIA OpenShell, a runtime that sandboxes agents using native Windows primitives. The system enables parallel execution of frontier agents and complements NVIDIA's physical AI agent skills release for high-throughput tasks. Hardware from partners like Dell and HP is expected to ship in Q4 2026 with 748GB of memory and 800Gb/s networking.
Still wondering? A few quick answers below.
Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →





